TLDR:
- I just watched the Neo trailer. It's creepy as hell, but I'm also really excited.
- The hardware design is smart: soft plastics, wool, and non-intimidating. It's designed to not be scary, even though it is.
- VR devices unlock cheap and accurate telerobotics which bridges the AI gap and gives data for building the AI
- This isn't just about chores; it's a "solid platinum" data collection play. They're getting priceless, labeled, in-home data to train their AI.
- It'll be built on exploited labor (like remote call centers), which is a problem, but it's the right, if imperfect, approach to getting Rosie the Robot who herself has problematic roots.
- This is a politics issue we need to tackle: When humans are no longer the primary productive unit our economic models fail or eliminate us
Here's the trailer: Neo - The First Humanoid Assistant
I recommend reading this article but if your too impatient and just want the raw stream of concioussness dump I edited it from in video form. Here it is: Comments on Neo
I just watched the Neo trailer. Neo is a new home robotics company that is launching a humanoid assistant to help you in your home. And let's be honest, the creep vibes are high and realistically it feels like the start of a horror movie. It's kind of scary.
But I'm also really excited and I do want to share my understanding of the issue and where I see it going. Ultimately, Neo, under the product of our current politics and our current capitalist system is going to be exploitative and in some ways deceptive at launch. But I really do think it is the right approach. I do think it signals a much better and hopeful future. And I want to explain my reasoning behind that.
The 'Creep Factor' is a Feature, Not a Bug
First off, Neo, as a product, appears to be engineered correctly. They have used a lot of soft plastics like TPU in its form factor. They have a wool overcoat design. The robot isn't overly strong. It's somewhat diminutive, but big enough to accomplish tasks as an adult. And they've just really headed out on the design past. Instead of entirely plastics, as I say, they're also using some what looks like cardboard or wood-derived materials. And the product overall, from a design point of view, is very friendly and not intimidating.
Because let's be real, they know the optics. Even in their own trailer, when people are greeting the robot for the first time and it emerges from this pod and asks you for your name, that feels like the start of a horror movie. We're culturally keyed to those indicators. And that's not a bad thing because our creatives in our society have been thought-exploring the space as philosophers and artists do to find the problems.
And if we hadn't been exploring this place, if we entered it without a degree of caution, that would be a problem. So it's not a bad thing that you feel creeped out. It's a good thing. It's a feature of society and it's something we should embrace. But that doesn't mean we should kill a potential future.
The Two-Part Problem: Hardware vs. Intelligence
Realistically, home robotics faces two problems. The problem of robotics and the problem of intelligence.
The problem of robotics is a hard one, but it's one we've been making iterative progress on and feels like we're pretty much there for entry level. We have a lot of problems still. This will have jank in places you don't expect and it is significantly weaker than a fit adult human. That's not a problem though. You don't need human matched strength and capabilities to do a few chores around the house. And from the intimidation factor, we probably want something that we can physically overpower. So, good design choices there.
Likewise, it has to deal with just the realities of moisture when doing things like the dishes and not taken care of when its owners are negligent. There are some very robust design challenges there. Will it hit everyone at launch? Probably not. But generally, people agree we're at the point where we can do that.
The other side is the battery life. The battery life is on the smaller side, but it seems more than sufficient. They rated it a few hours. I think it was up to four. But realistically, we know it's probably going to be in a range of two hours once you drain the battery a bit and go through a few cycles. But the fact that it can charge itself in, the fact that it has all day to take care of just your house means that it's very likely to be able to achieve that functionality.
So yeah, the physical side feels, if not solved, at the point where it is now useful. And it is very useful, especially when we start to talk about the elderly, especially when we talk about people who need home care needs. Can you imagine on the NHS being sent home from hospital and having an NHS robot assistant at your home for a short period?
The Telerobotics Genius
And that's where we get to the one that I think people will have the biggest challenge with, which is intelligence, telerobotics, and having someone in your home.
Now, this is a privilege of having growing up in South Africa... I grew up with maids and servants. And the reality is, there is a degree of trust there. Now, yes, when you hire someone, you have vetted them personally, but you give a lot of degree of trust to those people in your home. But on the other side... that's a giant ass robot, it's not going to sneak around. And I think we have to just accept if you're going to want this home assistance, there is going to be a robot in your home.
On the intelligence side, the reality is that for a long time, it is likely going to require cloud based or internet enabled intelligence. That being said, internet enabled and cloud based intelligence is not enough. So what they have done very smartly, is leaned into telerobotics.
Now, this would not have been possible without the VR revolution we've had, the affordability of a cheap quest or similar, XR2 powered device and 6DOF tracking really is a game changer for telerobotics.
Anyone who's worked in 3D software and VR just knows how much 6DOF controllers unlock. And previously manipulating a robotic arm through telerobotics required very expensive hardware that was rather clunky... And this rollout of VR has meant that we have all of these cheap, well-designed, heavily optimized telerobotics devices that are perfect for the task, at least from the torso up.
The difference of having a physical person, have to come to your home... versus someone remotely being able to do this work and only needing to cover the bits the robot cannot do, pretty huge.
Story Time: The Scoble Situation and Why Optics Matter
The other part is we want this to be monitored. Now we know from an optics point of view, the reality is at some point someone is going to do something criminal or harmful through telerobotics... It's inevitable. And I guarantee you every company that is operating in this space is stressing about it.
When we were working on the new revolution consumer VR headsets, a constant conversation was who was going to have the first death in VR... The reality is by the laws of averages, someone was going to have a heart attack... At some point, someone would die in a headset. And the horror was, if this happened close to launch... the tabloids could pick it up, and we could have a Scoble situation.
Richard Scoble, I think this is his name, unimportant. He's a journalist... a fat, white, pasty guy, body hair, who decided for internet points, he was going to shower with his Google Glass on. And the infamous picture of this pasty, overweight, white, middle-aged journalist, nerd, in your shiny new wearable, in the shower, tanked the public perception.
It was already going to be a hard challenge. And the glasshole adjective was firmly wedged into public consciousness. It was a huge deal and a massive lesson to tech... wearables are a fashion problem.
And you can clearly see with this robot, they have paid attention to that. The fact that it has the wool over body, the fact that the shoes are very similar to a ?vessi? design, the materials chosen, the fact it arrives in what looks like a recycled cardboard faux wood material. These are all product material design and fashion and aesthetic choices to put you at ease. And they're good choices.
The Exploitation and the 'Solid Platinum' Play
But it doesn't solve the problem that at some point, someone using a telerobotic setup... because let's be honest, call centers are going to go from people wearing headsets with phones on to people wearing VR headsets, right? Whether it's in a large scale call center in India, because it's going to happen like that, or probably not India, because India's economy is doing really well. They are going to find enough cheap labor in a place like America or the UK, because our economies aren't doing that great and it is better PR. Maybe it's going to be at home labor, which then you even lose another degree of oversight. But oversight is the key. This is monitored telerobotics.
And yes, while it may feel creepy that a company is recording things going on in your home, it's probably the safer option initially, because two reasons.
- When something inevitably does go wrong, you have that record of it.
- The reality is that if these employees are being exploited, which the reality is, they probably are going to be exploited low paid labor... The reality is they're not going to have power to turn off the monitoring, which is a good thing. I know it sounds terrible when I say it all like that.
And also the nature that those recordings exist, unlike a police officer who can turn off his body cam because he's like, "hmm, I'm going to do a little bit of naughty."
Now, if companies are choosing not to keep records, I'm even more concerned because they're probably thinking of the legal liability in the court case. And they're probably wargaming the situation where they're like, we don't want to actually have records because that makes us open to freedom of information requests and GDPR and all kinds of other stuff, which worries me more, actually. So yeah, no, hopefully there are recordings that can be used to resolve that.
But then the other slice, the other part of this pie is that we know intelligence, the biggest problem is data, good labeled data. And if you have even 100 of these robots out... let's say 90% of the time operated by a teller operator.
That's all labeled in world on hardware specific data, backed not only by the hand movements of the operator, but also all of the senses of the robot. And then probably even feedback from the owner... That kind of data set is... it's beyond gold dust. It's solid platinum in the current world.
Not only can you use it to improve your product, but... it is probably invaluable enough that it justifies the very existence of your company at this point to get in-home data like that. And that fidelity, I mean, every military corporate AI place wants that kind of scale of data. ...even if it's only used by the company itself to improve their robots... the reality is it will improve it, and it'll improve it fast.
This is How You Get Rosie the Robot
And the scary thing about AI that we always forget is... what I used to say, and I still say about VR, VR is the least comfortable it'll ever be today. It only gets better. And robotics and intelligence are the same.
Which we really need to struggle with on a politics and policy front. Because we really need to start re-evaluating a world where the unit of production is no longer a human. And if the unit of production is no longer a human and your entire political structure is based around productivity and production, the incentive models do not look good for us squishy soft people.
So that's a whole, but that's a politics issue. And it's one we should face up to. But ultimately, I'm excited. I'm really excited for Rosie the robot to become a reality. But we need to remember that Rosie the robot, when written, was a projection into the future of people in the present were being exploited. She's clearly modeled after, well, because it was an American cartoon, the African American servant or maid. She was clearly modeled after slavery in many ways, which is an awkward factor of our culture we have to like line up with.
That is, I think, more a projection of the period it was made in and the art. The reality is home robots, retail robots, commercial robots are a huge unlock for the next stage of human growth. It's a big deal. And it's something we should be excited about... even if the road is imperfect. Though wealth inequiality is a CORE ISSUE around robotics.
But don't be blind. There's a reason the movies and media has warned us about robots. Because while we should move forward, if we take our eye off the road, it's very possible we careen off it.
I AM STILL EXCITED FOR FREKKIN ROBOTS!!!
But no, I'm not going to be able to afford that thing. It's fucking expensive as hell.